Shaping the Neocortex: Radial Glia and Astrocytes in Development and Evolution
- Federica Mosti1,2,
- Hiroshi Kawasaki3,
- Courtney Babbit4,
- Barbara di Benedetto5,6,
- Debra L. Silver1,2,
- Nereo Kalebic7, and
- Carmen Falcone8
- 1Departments of Molecular Genetics and Microbiology, Durham, North Carolina 27710
- 2Neurobiology, Duke University Medical Center, Durham, North Carolina 27710
- 3Department of Medical Neuroscience, Graduate School of Medical Sciences, Kanazawa University, Kanazawa, Ishikawa 920-8641, Japan
- 4Department of Biology, UMass Amherst, Amherst, Massachusetts 01003
- 5Department of Psychiatry and Psychotherapy, University of Regensburg, Regensburg 93053, Germany
- 6Regensburg Center of Neuroscience, University of Regensburg, Regensburg 93053, Germany
- 7Human Technopole, Milan 20157, Italy
- 8Department of Biological Sciences, Towson University, Towson, Maryland 21252
-
Author contributions: F.M., H.K., C.B., B.B., D.L.S., N.K., and C.F. wrote the paper.
Abstract
The evolutionary expansion of the mammalian neocortex—especially in primates—underpins the emergence of advanced cognitive abilities. This process involved not only increased cortical surface area and neuronal output but also enhanced structural adaptations, such as cortical folding and glial morphological complexity. In this review, we examine the central roles of radial glia (RG) and astrocytes in driving neocortical expansion and evolution. We highlight the emergence of primate- and human-specific genes, which contribute to enhanced RG proliferation and neurogenesis in these species. We further explore how epigenetic regulation and dynamic chromatin architecture modulate RG behavior across species. At the cellular level, we discuss how morphological features—particularly the basal processes and specialized protrusions of RG—facilitate access to diverse extrinsic signals, promoting proliferative capacity and cortical complexity. We then turn to cortical folding, focusing on the role of astrocytes, and the functional relevance of folds in supporting brain homeostasis. Finally, we address astrocyte diversity, development, and evolutionary adaptation, with special emphasis on sex differences and primate-specific features. Comparative transcriptomic and morphological studies reveal that human astrocytes exhibit unique molecular signatures, expanded metabolic capacity, and higher morphological complexity. Together, these insights underscore the multifaceted contributions of RG and astrocytes to the evolutionary elaboration of the neocortex. They further provide a framework for understanding how cellular innovations shaped the modern primate brain in general, and human brain specifically.
Introduction
The neocortex is the largest part of the cerebral cortex and is involved in higher cognitive functions in mammals (Rakic, 2009; Sousa et al., 2017; Galakhova et al., 2022). Throughout mammalian evolution, the neocortex has undergone significant enlargement in many species, leading to considerable variations in its size. Although this evolutionary expansion occurred independently in most mammalian orders (Lewitus et al., 2014; Del-Valle-Anton and Borrell, 2022), it is believed to follow common underlying principles. The enlargement of the neocortex involves increases in both surface area and thickness (Rakic, 2009), with the most pronounced expansion of surface area. To accommodate the increased number of cells across the expanded cortical surface, cortical folding emerged as a key evolutionary adaptation (Molnár et al., 2006; Rakic, 2009; Kelava et al., 2013; Zilles et al., 2013; Sun and Hevner, 2014; Llinares-Benadero and Borrell, 2019).
The main parameters thought to underlie neocortical expansion include the following: (1) the rate at which neural progenitor cells produce differentiated cells (neurons and glia); (2) the duration of neocortical development; and (3) the survival rate of these cells (Lui et al., 2011; Florio and Huttner, 2014; Del-Valle-Anton and Borrell, 2022).
Neural progenitor cells are typically classified into radial glia (RGs) and intermediate progenitor cells (IPCs; Fig. 1A). RG possess a higher proliferative capacity, giving rise to additional RG, IPCs, and differentiated cells, whereas IPCs are more lineage restricted and primarily generate differentiated cells directly (Taverna et al., 2014). Based on the site of mitosis and contact with the apical surface, progenitor cells are further categorized into apical and basal progenitors (Taverna et al., 2014). Accordingly, RG are divided into two subtypes: apical radial glia (aRG) and basal or outer radial glia (bRG or oRG). While aRG are abundant and proliferative across all mammals, bRG exhibit striking interspecies differences that correlate with neocortical expansion (Lui et al., 2011; Florio and Huttner, 2014).
Key cellular players and structural features shaping neocortical development and evolution. A, Schematic of neurogenesis and gliogenesis in the developing neocortex, depicting the lineage progression from neuroepithelial cells to apical radial glia (aRG), intermediate progenitors (IP), and basal radial glia (bRG). Differentiated neurons (deep and upper layer) and astrocytes arise from these progenitors. B, Comparative organization of the neocortex in mouse, ferret, and human, highlighting increased progenitor diversity and laminar complexity in species with larger, folded cortices. C, Functional roles of RG basal processes during development, including scaffolding for neuronal migration, signaling interactions with the basal lamina and meninges, and subcellular compartmentalization. D, Overview of astrocyte functions relevant to neocortical development and evolution, including regulation of synaptic transmission, metabolic support of neurons, blood–brain barrier maintenance, and astrocyte–astrocyte and microglia interactions.
In species with an enlarged and folded neocortex, bRG are more numerous and possess greater proliferative capacity (Kalebic and Huttner, 2020; Fig. 1B). Since aRG are restricted to the ventricular zone (VZ), their expansion primarily contributes to the lateral growth of the neocortex (Rakic, 1988, 2009). In contrast, bRG (Fig. 1C) acquired the ability to proliferate away from the ventricular surface, within the subventricular zone (SVZ). Their increased abundance leads to a pronounced thickening of the SVZ, which, in species with an expanded neocortex, becomes subdivided into inner and outer SVZ (iSVZ and oSVZ, respectively; Dehay et al., 2015; Fig. 1B). This expansion of the bRG pool results in a dramatic increase in the production of differentiated cells, leading in turn to a larger and more complex neocortex (Kalebic and Huttner, 2020).
During development, RG initially give rise to neurons, and subsequently generate glial cells, such as astrocytes and oligodendrocytes. Astrocytes are a diverse and dynamic population of glial cells playing essential roles within the central nervous system (CNS; Fig. 1D). They regulate key processes including neural development, brain homeostasis, blood–brain barrier formation and maintenance, and synaptic transmission and plasticity (Verkhratsky and Nedergaard, 2018; Tan et al., 2021; Lyon and Allen, 2022; Lawal et al., 2022; Verkhratsky and Butt, 2023; Raghunathan and Eroglu, 2025). Astrocytes are characterized by remarkable morphological and functional heterogeneity, even within a single species (Matyash and Kettenmann, 2010; García-Marqués and López-Mascaraque, 2013; Sosunov et al., 2014; Köhler et al., 2019; Yang and Jackson, 2019; Batiuk et al., 2020; Bayraktar et al., 2020). In the cerebral cortex, three major anatomical types of astrocytes can be distinguished: protoplasmic, fibrous, and interlaminar. Protoplasmic astrocytes populate the gray matter and exhibit complex arborizations closely associated with numerous synapses, whereas fibrous astrocytes primarily reside in the white matter and have elongated, less branched processes (Verkhratsky and Butt, 2023). Interlaminar astrocytes, on the other hand, have their cell bodies in layer I, close to the pial surface, and extend distinctively long processes that cross multiple cortical layers (Colombo et al., 2000; Colombo and Reisin, 2004; Falcone et al., 2019; Ciani and Falcone, 2024). Recent advancements, particularly in single-cell RNA sequencing (scRNA-Seq) and spatial transcriptomics, have revealed even finer astrocyte diversity, showing regional specialization (Batiuk et al., 2020; Bayraktar et al., 2020) and distinct functional roles within neural circuits (Lyon and Allen, 2022). These findings emphasize that astrocyte identity is shaped by both intrinsic developmental programs and by local environmental cues, including neuronal activity (Morel et al., 2014; Hasel et al., 2017; Farhy-Tselnicker et al., 2021).
When considering astrocytes across different mammalian species, this heterogeneity becomes even more pronounced. Astrocytes have progressively diversified in parallel with increasing brain size and complexity (Falcone, 2022). Primate astrocytes show particularly sophisticated morphological and molecular adaptations (Oberheim et al., 2009; Falcone et al., 2019; Falcone, 2022; Ciuba et al., 2025). They are significantly larger, structurally more complex, and capable of interacting with a larger number of synapses, compared with astrocytes from other mammals (Oberheim et al., 2009). Moreover, certain astrocytic subtypes, such as interlaminar astrocytes, show primate-specific features, suggesting that these cells play a critical role in the evolution of cortical organization and function (Falcone et al., 2019, 2020).
In addition to evolutionary and species-specific mechanisms, intrinsic biological variables such as sex significantly influence brain development. From early gestation through postnatal life, male and female brains diverge in terms of developmental timing, regional growth patterns, synaptogenesis, and the emergence of neural circuits. These differences are not solely driven by gonadal hormones but also involve gene expression regulation and epigenetic mechanisms operating in a sex-specific manner. Glial cells, and astrocytes in particular, have emerged as key mediators of these sexually dimorphic processes, influencing everything from synaptic pruning to the maturation of neural connectivity (Peper et al., 2020; Bruno et al., 2022; Casado-Navarro and Serrano-Saiz, 2022; Freire-Regatillo et al., 2022; Kim and Kim, 2022; Lagunas et al., 2022). Recognizing sex as a biological variable is thus crucial to understanding the cellular and molecular diversity underlying brain development and vulnerability to neurodevelopmental disorders.
In this review, we explore the critical contributions of RG and astrocytes to neocortical development and discuss how their specializations have facilitated the evolutionary expansion and increased complexity of mammalian brains, specifically primates. We highlight unique molecular profiles, regulatory mechanisms, morphological innovations, and metabolic adaptations in both RG and astrocytes across species. We further discuss how molecular and cellular features of RG and astrocytes have supported the remarkable expansion of primate neocortex which is at the basis of our cognitive abilities. By integrating recent findings obtained through multiple methodologies and across different species, this review aims to highlight the central role of both RG and astrocytes in shaping the evolution of the mammalian neocortex.
Human- and Primate-Specific Genes in RG Biology
In mice, bRG are relatively rare and exhibit low proliferative capacity (Wang et al., 2011; Wong et al., 2015) whereas in species with a more complex and expanded neocortex—such as ferrets, macaques, and humans—they are significantly more abundant and demonstrate higher proliferative potential (Fietz et al., 2010; Hansen et al., 2010; Reillo et al., 2011; Betizeau et al., 2013; LaMonica et al., 2013; Kalebic et al., 2019). To elucidate the mechanisms underlying this enhanced proliferative ability in gyrencephalic species (i.e., characterized by a highly folded cerebral cortex, marked by the presence of ridges called gyri and grooves known as sulci), recent research has focused on both evolutionary emergence of novel genes and differences in gene expression regulation across species.
Several human- and primate-specific genes expressed by RG were recently shown to play a role in the evolutionary expansion of the primate neocortex. ARHGAP11B is a hominin-specific gene that arose ∼5 million years ago via partial duplication of ARHGAP11A and acquired a unique point mutation enabling it to expand bRG pool and promote neocortex enlargement (Florio and Huttner, 2014; Florio et al., 2016). ARHGAP11B is highly expressed in neural stem and progenitor cells of the fetal neocortex—specifically in aRG of the ventricular zone and bRG of the SVZ (Florio et al., 2015). Its introduction in mouse, ferret, and marmoset neocortex in vivo, as well as chimpanzee organoids, increases basal progenitor abundance and promotes and extends neurogenesis, thus leading to cortical folding (Florio and Huttner, 2014; Kalebic et al., 2018; Heide et al., 2020; Fischer et al., 2022). ARHGAP11B transgenic mice exhibit greater memory flexibility, suggesting a link with enhanced cognitive abilities acquired during human evolution (Xing et al., 2021). Mechanistically, ARHGAP11B enhances glutaminolysis (Namba et al., 2020), suggesting metabolic adaptation as an evolutionary mechanism. NOTCH2NL is a human-specific gene family that arose from partial duplication and gene conversion of the conserved NOTCH2 gene. NOTCH2NL is mainly expressed in the germinal zones of the developing cortex, with notable enrichment in aRG and bRG (Fiddes et al., 2018; Suzuki et al., 2018). Functional studies in mice and human organoids show that NOTCH2NL delays neurogenesis and maintains RG pool through activation of Notch signaling, while its loss accelerates differentiation (Fiddes et al., 2018; Suzuki et al., 2018). Finally, recently identified human-specific NBPF14 was shown to increase bRG pool by promoting delamination of aRG (Eşiyok et al., 2025).
TBC1D3 and TMEM14B are primate-specific genes—both expressed in RG during development—that promote neocortical expansion by increasing bRG abundance. TBC1D3, which arose from segmental duplication of USP6NL and expanded in copy number in humans, induces aRG delamination and bRG proliferation, leading to cortical folding (Hodzic et al., 2006; Ju et al., 2016). TMEM14B, a probable paralog of TMEM14C, is conserved in apes and Old World monkeys and promotes both generation and expansion of bRG, leading to an enlargement and increased folding of the neocortex (Liu et al., 2017).
Although many human and primate-specific genes have played a role in neocortical expansion, as discussed above, it is important to consider that neocortical size has progressively increased throughout mammalian evolution. This suggests that also conserved mechanisms are at play in regulating neocortical expansion. One example is the extracellular signal-regulated kinase (ERK)-bone morphogenic protein 7 (BMP7)-GLI3 repressor form (GLI3R)-sonic hedgehog (SHH) positive feedback loop (Yang, 2025). In summary, cortical expansion likely relies on a combination of evolutionary genetic adaptations unique to certain lineages and fundamental developmental mechanisms shared across species.
While the emergence of human- and primate-specific genes has played a role in enhancing RG proliferation and neocortical expansion, these genetic innovations function within a broader regulatory landscape. Indeed, beyond the appearance of novel genes, evolutionary changes in gene regulation—mediated by DNA methylation, histone modifications, and chromatin architecture—are essential for orchestrating the timing, location, and type of RG behavior. Such epigenetic mechanisms have a pivotal role in fine-tuning gene accessibility and responsiveness during cortical development. The following section explores how these epigenetic processes dynamically shape RG identity and output, offering further insight into the mechanisms that underlie human brain development and evolution.
Epigenetic Regulation of RG
DNA methylation and histone modifications are key epigenetic modulators which reorganize chromatin to control accessibility of transcriptional machinery (Lewis et al., 2021). These modifications shape the transcriptional landscape of RG, enabling precise control over when and how these progenitors proliferate and differentiate (Roadmap Epigenomics Consortiumet al., 2015). One of the pivotal epigenetic mechanisms in RG is DNA methylation, predominantly orchestrated by DNA methyltransferases (DNMTs). During corticogenesis, DNMT1 is continuously expressed, with levels increasing from progenitor to postmitotic neuron stages (Sanosaka et al., 2017). This tightly regulated methylation control is highly dynamic and cell-type specific, ensuring proper lineage specification. DNMT1 maintains the suppression of gliogenic genes during early neurogenesis, thereby preserving the neurogenic potential of RG (Sanosaka et al., 2017). Notably, loss of DNMT1 leads to premature astrocyte differentiation and severe cortical malformations, highlighting its essential role in sustaining normal cortical development (Amberg et al., 2019).
Histone modifications also play a decisive role in determining RG fate. The Polycomb Repressive Complex 2 (PRC2), via its catalytic subunit EZH2, deposits the repressive H3K27me3 marker to silence neurogenic genes (Bölicke and Albert, 2022). Disruption of EZH2 function results in premature differentiation of RG, excessive production of deep-layer neurons, depletion of upper-layer neurons, and early onset of gliogenesis (Pereira et al., 2010; Hahn et al., 2013). In contrast, the histone methyltransferase PRDM16 activates enhancers crucial for IPC production and proper cortical lamination (Baizabal et al., 2018). These contrasting roles underscore how histone modifiers coordinate timing and identity of neural output. Importantly, mutations in epigenetic regulators are implicated in intellectual disabilities and cortical malformations (Pereira et al., 2010; Hahn et al., 2013; Baizabal et al., 2018), underscoring the delicate epigenetic balance required for healthy brain development.
Beyond biochemical modifications, chromatin architecture itself is a powerful regulator of RG function. The genome is folded into a three-dimensional structure that evolves alongside developmental cues (Harabula and Pombo, 2021). In early corticogenesis, RG chromatin is more open and accessible, enabling rapid transcriptional responses. As development progresses, chromatin compacts and accumulates repressive histone marks, limiting transcriptional flexibility (Bonev et al., 2017). This dynamic regulation of chromatin states facilitates RG transitions from multipotency to lineage commitment. When these processes go awry, they can disrupt nuclear architecture and gene expression programs, which are hypothesized to contribute to the pathogenesis of neuropsychiatric disorders (Harabula and Pombo, 2021). 3D chromosome interaction maps have been generated across diverse species (Won et al., 2016; Bonev et al., 2017; Luo et al., 2021; Keough et al., 2023), though the evolutionary relevance is still unknown.
The reconfiguration of enhancer–promoter interactions is central to the emergence of human-specific complex brain functions. Enhancers and promoters tightly orchestrate spatiotemporal gene expression in RG (Amiri et al., 2018). Human-specific regulatory elements, such as human accelerated regions (HARs) and human accelerated quantitative enhancer regions (HAQERs), have evolved rapidly and enriched nearby neurodevelopmental genes (Franchini and Pollard, 2017; Mangan et al., 2022). The active role of these evolutionarily unique regions in shaping the human brain has recently gained attention. For example, HAR enhancer 5 (HARE5) has been shown through genome-edited primate and mouse models to regulate cortical growth by influencing RG proliferation and neurogenesis (Boyd et al., 2015; Liu et al., 2025). Furthermore, genome-wide association studies (GWAS) have identified significant enrichment of autism spectrum disorder (ASD) and schizophrenia-associated variants within HARs (Xu et al., 2015; Doan et al., 2016). This suggests that evolutionary innovations in gene regulation may also contribute to disease susceptibility.
bRG Morphology in the Neocortical Evolution
At the cell biological level, various genetic and epigenetic mechanisms described above are integrated in different aspects of RG biology, including cell morphology. While bRG were initially characterized as monopolar cells with a single basal process extending to the pial surface (Fietz et al., 2010; Hansen et al., 2010; Reillo et al., 2011), later studies in macaque, ferret, and human reported a greater bRG morphological heterogeneity (Betizeau et al., 2013; Reillo et al., 2017; Kalebic et al., 2019; Fig. 2). Interestingly, bRG characterized by multiple cellular processes are associated with higher proliferative capacity (Kalebic and Huttner, 2020). Notably, bRG and basal IPCs in the human and ferret neocortex have more processes than those in the mouse (Kalebic et al., 2019; Fig. 2). The identification of morphoregulatory proteins such as PALMD and adducins has enabled the experimental induction of additional processes in mouse and ferret bRG, resulting in increased proliferation. Conversely, reducing the number of bRG processes in human fetal neocortical tissue ex vivo leads to decreased proliferation, underscoring the functional link between bRG morphology and proliferative potential (Kalebic et al., 2019).
Morphological heterogeneity of basal radial glia and astrocyte subtypes across species. Classification of bRG morphotypes and astrocyte subtypes based on their morphological features in ferret, mouse, and human developing neocortex. Comparative morphology of bRG and astrocytes highlighting species-specific differences in process complexity and density. Human bRG and astrocytes exhibit more elaborate morphologies than those in mouse, consistent with enhanced proliferative and functional capacity.
The basal process of bRG plays a central and multifaceted role in neocortical development and evolution (Taverna et al., 2014; Kalebic and Namba, 2021; Viola et al., 2024; Fig. 1C). It serves two key functions: (1) as a scaffold for radial neuronal migration toward the cortical plate and (2) as a structural and functional extension that supports progenitor proliferation (Kalebic and Huttner, 2020). The long basal process enables bRG to maintain contact with distant structures such as the basal lamina and to interact with a variety of cortical layers along its length, potentially integrating diverse signals. In addition to their long radial processes, some RG develop small lamellate expansions, which increase during development and may facilitate signal reception from the surrounding tissue—a feature observed in primates and ferrets but not in mice (Rakic, 1972; Schmechel and Rakic, 1979; Kalebic et al., 2019). Together, these data suggest that cellular protrusions expose RG to diverse and enhanced pro-proliferative signaling from the environment. Multiple extracellular signals, including growth factors such as PDGF (Lui et al., 2014), FGF (Rash et al., 2013; Matsumoto et al., 2017)—in particular the FGF-FGFRs-pERK axis (Sun et al., 2024)—and Epiregulin (Cubillos et al., 2024); morphogens such as SHH (Wang et al., 2016; Matsumoto et al., 2020), WNT (Kuwahara et al., 2010), and BMP (Najas et al., 2020; Li et al., 2024); neurotransmitters such as serotonin (Xing et al., 2020); along with extracellular matrix components (Fietz et al., 2012; Pollen et al., 2015; Buchsbaum et al., 2020; Güven et al., 2020), Notch ligands (Hansen et al., 2010), and others, have been identified and shown to play a role in RG proliferation (Ferent et al., 2020; Kalebic and Huttner, 2020).
While the role of the bRG basal process in subcellular compartmentalization is less understood, parallels with aRG are emerging (D’Arcy and Silver, 2020). In aRG, the basal process and end foot are sites of localized translation, containing specific mRNAs and RNA-binding proteins, including those for key cell cycle regulators like cyclin D2 (Tsunekawa et al., 2012; Pilaz et al., 2016, 2023; D’Arcy et al., 2023; Fig. 1C). These findings underscore the need for further investigation into the compartmentalization and trafficking dynamics within the bRG basal process.
Altogether, these findings highlight the integral role of bRG morphology in shaping neocortical development and evolution (Fig. 2). Understanding how morphological specializations modulate cellular behaviors will be instrumental to uncover the mechanisms underlying evolutionary expansion of the neocortex.
RG and Cortical Folding
As mentioned above, the increased production of neurons and glia during mammalian brain evolution led to the expansion of the neocortex. This was accompanied by the appearance of cortical gyri and sulci in various mammalian orders (Molnár et al., 2006; Rakic, 2009; Zilles et al., 2013; Sun and Hevner, 2014; Llinares-Benadero and Borrell, 2019). It is believed that these evolutionary changes are linked to higher brain functions, as abnormalities in cortical folding are associated with severe cognitive dysfunctions in humans. Nevertheless, the mechanisms underlying the development and evolution of cortical folds remain poorly understood. One reason for this is that mice—the most common model in genetic studies—have small, smooth brains and thus limited utility in studying cortical folding. Ferrets, with their larger, gyrencephalic brains, have thus emerged as valuable models (Gilardi and Kalebic, 2021; Kawasaki, 2024). The development of genetic manipulation tools in ferrets has further enabled the study of molecular mechanisms underlying cortical development and evolution (Kawasaki et al., 2012).
Using these tools, it was shown that FGF and SHH signaling regulate the proliferation and differentiation of bRG, respectively, promoting the generation of upper-layer neurons (Matsumoto et al., 2017, 2020). This preferential expansion of upper layers initiates cortical folding during development (Shinmyo et al., 2017). In addition to increased neurogenesis, astrocyte proliferation also contributes to cortical folding (Rash et al., 2019). Specifically, a positive feedback loop involving FGF signaling in astrocytes sustains the continued growth of cortical folds after their initial formation (Shinmyo et al., 2022). These findings suggest that the synergistic interaction between neurogenesis and astrogenesis is essential for the formation of fully developed cortical folds during development.
The biological significance of cortical folds is still debated. Conventional hypotheses propose that folding enables more neurons to fit within the limited space of the skull (Molnár et al., 2006; Rakic, 2009; Zilles et al., 2013; Sun and Hevner, 2014; Llinares-Benadero and Borrell, 2019). However, the widespread presence of cortical folds in mammalian cerebrum suggests additional functional advantages. In this context, the glymphatic system has been studied. The glymphatic system is a brain-wide waste clearance pathway that facilitates the removal of interstitial solutes, including neurotoxic proteins such as amyloid-β, through cerebrospinal fluid (CSF) exchange with interstitial fluid (Mestre et al., 2020). This system relies on perivascular channels formed by astrocytic end feet expressing the water channel aquaporin-4 (AQP4), which line the vasculature and enable directional CSF flow through the parenchyma (Iliff et al., 2012; Hablitz and Nedergaard, 2021). Astrocytes are the central regulators of this process, as their polarized expression of AQP4 at the perivascular end feet is essential for the convective bulk flow that drives glymphatic clearance. Disruption of AQP4 localization or expression, as observed during aging or in disease states, significantly impairs glymphatic function and exacerbates the accumulation of metabolic waste (Kress et al., 2014; Ding et al., 2023). In addition to AQP4, astrocytes regulate glymphatic dynamics through their roles in ion homeostasis, vascular tone modulation, and reactivity in response to neuroinflammation. Recent work has further revealed that astrocyte-dependent glymphatic function is not uniform across the brain, varying across regions and being influenced by sex, age, and pathological context (Sriram et al., 2024). These findings highlight astrocytes as active gatekeepers of brain waste clearance and underscore their broader role in maintaining cerebral homeostasis and protecting against neurodegenerative processes.
While previous studies on the glymphatic system primarily used mouse models, its function in more complex, folded brains remained unclear. As the cerebrum expanded during evolution, its deeper regions became increasingly distant from the surface, potentially impeding CSF penetration and waste clearance. By studying the glymphatic system in ferrets, a novel circulation pattern unique to the gyrencephalic cerebrum was discovered (Kameya et al., 2024), providing compelling evidence that cortical folds enhance glymphatic circulation in enlarged mammalian brains.
These findings highlight the functional significance of astrocytes in mammalian brain evolution, suggesting a role for them not only in structural adaptations such as cortical folding but also in supporting essential physiological processes including glymphatic clearance. Astrocytes' broad contributions to cerebral architecture and homeostasis suggest that their diverse roles may be further nuanced by intrinsic biological factors. In the next sections, we delve into astrocyte development and how their intrinsic biological factors, such as sex differences, shape neuronal circuitry and have consequences on brain evolution and neurodevelopmental disorders.
Astrocyte Development: A Focus on Sex Differences
As discussed in the introduction, astrocytes play a central role in shaping the developmental landscape of the brain by regulating synaptogenesis, synaptic pruning, glutamate clearance overall circuit stabilization (Chung et al., 2015; Tan et al., 2021; Vivi and Di Benedetto, 2024), and formation and maintenance of the blood–brain barrier (BBB; Guérit et al., 2021; Manu et al., 2023; Fig. 1D). In this critical postnatal period, sex differences are increasingly recognized as key biological variables shaping brain development and long-term susceptibility to neurodevelopmental and psychiatric disorders (Sengpiel, 2007; Ziemka-Nalecz et al., 2023). Notably, astrocyte maturation follows sex-specific trajectories: in male rodents, astrocytes reach maturity earlier than in females (Rurak et al., 2021, 2022), a process largely driven by the perinatal surge in testosterone and its local conversion to estradiol via astrocyte-expressed aromatase, which acts to masculinize both neural and glial networks (Arnold and Gorski, 1984; Amateau and McCarthy, 2002; Bakker, 2019; Rurak et al., 2023). These temporal differences in astrocyte development have significant implications for the regulation of excitatory/inhibitory (E/I) balance (Ribot et al., 2021). Estrogenic modulation in females also influences astrocyte function in a region- and time-dependent manner (Ziemka-Nalecz et al., 2023). These hormonal effects are essential for the timely closure of developmental windows of plasticity; disruption or any premature or delayed closure may lead to sex-biased vulnerabilities. For instance, premature closure of astrocyte-mediated plasticity windows may lead to altered E/I ratios and can contribute to early-onset conditions such as ASD, which disproportionately affects males. Conversely, delayed astrocyte maturation in females may prolong periods of circuit plasticity, potentially increasing susceptibility to affective disorders emerging later in life, such as anxiety or mood disorders (Dehorter and Del Pino, 2020; Rurak et al., 2021; Collignon et al., 2024).
Rodent models have been instrumental in elucidating these sex-specific glial developmental trajectories, offering the advantage of well-characterized postnatal development and the ability to manipulate hormone levels and genetic factors. In these models, disruption of gonadal hormone signaling through neonatal castration, aromatase inhibition, or receptor blockage delays astrocyte maturation and reduces glutamate transporter expression in males, effectively feminizing glial developmental trajectories (Bakker, 2019; Collignon et al., 2024). However, rodent astrocytes differ significantly from human astrocytes in morphology, gene expression, and functional capacity, limiting the translatability of findings. Human-derived systems, such as induced pluripotent stem cell (iPSC)-derived astrocyte (iAstrocyte) cultures and brain organoids, are helping to close this gap, though they currently lack the full cellular diversity and long-range circuit complexity necessary to fully model E/I dynamics in vivo. Integrating data across different model systems—animals, in vitro, and in silico—will be essential for better understanding how sex-specific astrocyte maturation during postnatal development shapes neural circuit function and influences disease susceptibility. These insights may support the development of sex-informed therapeutic strategies for neurodevelopmental and psychiatric disorders.
In addition to sex-based differences during postnatal development, astrocytes exhibit evolutionary adaptations across species—particularly in primates—that reflect and support increased brain complexity and cognitive capacity. The following section explores how these evolutionary changes have shaped astrocyte form and function, highlighting structural innovations that distinguish primate astrocytes from their nonprimate counterparts.
Astrocyte Morphological Changes across Evolution
Diversity in astrocytes across mammals was first highlighted in the seminal anatomical studies by Ramón y Cajal (1909–1911). Specifically, astrocytes exhibit remarkable specializations in primates in terms of density, morphology, and molecular signatures (Falcone et al., 2019; Verkhratsky et al., 2019; Ciani and Falcone, 2024; Zintel et al., 2024; Ciuba et al., 2025) which likely contributed significantly to the evolution of primate brain complexity. Both protoplasmic and fibrous astrocytes in primates are larger and morphologically more elaborate than those in nonprimates. For instance, human protoplasmic astrocytes exhibit a diameter ∼2.6 times larger than their rodent counterparts and possess, on average, ∼10 times more primary processes (∼40 in humans vs 3–4 in rats; Oberheim et al., 2009).
Interlaminar astrocytes are an example of another impressive primate-specific specialization. These astrocytes are widely distributed across mammals, in either typical or rudimentary forms (Falcone et al., 2019); however, primate interlaminar astrocytes display distinctive features regarding density, morphology, molecular markers, and developmental trajectories (Colombo and Reisin, 2004; Falcone et al., 2019). While their cell bodies reside in cortical layer I near the pia, their processes extend perpendicularly, crossing multiple cortical layers (Colombo, 2017). Recent comparative studies across several mammalian orders revealed that interlaminar astrocytes are significantly more abundant and morphologically complex, characterized by longer processes, increased branching, and deeper penetration into the cortex in primates compared with nonprimates (Falcone et al., 2019; Falcone and Martínez-Cerdeño, 2023; Fig. 2). Additionally, primate interlaminar astrocytes exhibit a prolonged developmental trajectory extending months postnatally in macaques, chimpanzees, and humans. This extended maturation includes sustained expression of RG markers (such as PAX6, SOX2, NESTIN), indicating a potential RG origin. Furthermore, primate interlaminar astrocytes express molecular markers such as S100, CRYAB, and HOPX, which are absent in rodent layer I astrocytes, underscoring their unique primate molecular identity (Falcone et al., 2020).
Varicose projection astrocytes are another astrocyte subtype that has been initially well characterized in humans and other hominoids, including great apes; however, they have been recently found to be present in other mammals such as mice and tigers (Ciani et al., 2025). Varicose projection astrocytes reside predominantly in the deeper cortical layers (layers V–VI) and white matter. They are characterized by their peculiar morphology featuring one to five long processes exhibiting regularly spaced varicosities (Oberheim et al., 2009; Falcone et al., 2021). These processes extend tangentially across significant cortical distances, often traversing domains of neighboring astrocytes and interacting with blood vessels (Oberheim et al., 2009; Falcone et al., 2021). Interestingly, varicose projection astrocytes show substantial interindividual variability, suggesting a reactive phenotype triggered by specific conditions rather than constituting a stable, physiological astrocytic subtype. This hypothesis is supported by their irregular occurrence and coappearance with varicosities in interlaminar astrocytes under similar conditions, potentially indicating shared underlying cellular responses (Falcone et al., 2021).
Intriguingly, the ferret also shows multiple astrocyte subtypes in both gray and white matter that resemble those found in primates. In cortical layer I, ferrets have not only protoplasmic astrocytes but also interlaminar astrocytes. In the white matter, four distinct astrocyte subtypes were distinguished, classified by their process density and presence of varicosities (Tai et al., 2024). Additionally, ferret astrocytes exhibit significantly larger territorial domains and process lengths compared with mouse astrocytes (Tai et al., 2024; Fig. 2).
Human astrocytes also demonstrate unique molecular and functional adaptations compared with other primates. Recent studies using iAstrocytes from humans, chimpanzees, and macaques revealed that human iAstrocytes are larger and structurally more complex than their nonhuman primate counterparts, reflecting evolutionary enhancements. Human astrocytes exhibit enhanced long-range intercellular signaling and reduced expression of neurological disorder-related genes compared with chimpanzee and macaque astrocytes, suggesting molecular changes associated with brain evolution. These molecular distinctions are attributed to evolutionary gains in regulatory enhancer regions and transcription factor binding sites that contribute significantly to human-specific astrocytic complexity and function (Ciuba et al., 2025).
Evolutionary Differences in Astrocyte Gene Expression
Comparative transcriptomic studies have been highly valuable to examine such genomic changes across astrocyte evolution. We are now beginning to understand more about the functional changes in astrocyte evolution by looking at gene expression in tissue, astrocytes in culture, and organoids. Combining such advanced model systems with scRNA-seq has highlighted gene expression profiles that distinguish human astrocytes from those of other primates (Micali et al., 2023; Ciuba et al., 2025). These can at the basis of cellular and functional differences which contribute uniquely to the complexity of the human brain. For example, human astrocytes have been shown to have a higher density, soma volume, and ratio of GFAP+ astrocytes to total glial cells in layer I in humans compared with other primates (Munger et al., 2022). Understanding these species-specific differences within primates, especially comparing humans and chimpanzees, is vital to understanding the underpinnings of our metabolically demanding brain.
Comparative studies between humans and nonhuman primates have shown that gene expression diverged more rapidly in glial cells compared with neurons (Berto et al., 2019; Jorstad et al., 2023). In particular, human astrocytes showed the highest number of differentially expressed genes, with many enriched in synaptic signaling, protein translation, and cell adhesion functions. A recent single-nuclei RNA-seq study observed a substantial increase in transcriptomic heterogeneity in rhesus macaque astrocytes compared with marmoset astrocytes (Ma et al., 2022). Human iAstrocytes show twice as many differentially expressed genes relative to chimpanzee iAstrocytes than human neurons do compared with chimpanzee neurons (Zintel et al., 2024). Furthermore, human iAstrocytes show a greater activation of genes and pathways implicated in long-range intercellular signaling compared with chimpanzee and macaque (Ciuba et al., 2025). Moreover, genes frequently associated with neurological disorders are downregulated in human iAstrocytes compared with other primates (Ciuba et al., 2025). These studies found significant overrepresentation of genes related to metabolism which are differentially expressed in astrocytes across species (Zintel et al., 2024; Ciuba et al., 2025). Organoid studies combined with scRNA-Seq have also provided important insights into astrocyte function, development, and maturation (Zhang et al., 2016; Sloan et al., 2017; Kanton et al., 2019; Velasco et al., 2019) and have revealed species-specific differences in astrocyte gene expression (Kanton et al., 2019; Agoglia et al., 2021).
Hybrid fused human–chimpanzee tetraploid cells are also a useful model to look at gene expression from an evolutionary perspective (Gokhman et al., 2021; Song et al., 2021). Human and chimpanzee iPSCs are first fused to generate tetraploid hybrid stem cells and then differentiated into neural organoids. These models showed significantly higher expression of the human alleles in the astrocyte gene set module, indicating lineage-specific selection for either upregulation in humans or downregulation in chimpanzees (Agoglia et al., 2021). This gene set module was enriched for astrocyte-associated genes (e.g., AQP4 and S100) which may have evolved to enable adaptive species-specific astrocyte functionality (Agoglia et al., 2021).
Altogether, the transcriptomic studies discussed above have revealed extensive molecular adaptations that distinguish human astrocytes from those of other primates and suggest evolutionary changes in metabolism, signaling, and cell adhesion. These molecular shifts likely underpin broader morphological adaptations, given the close relationship between gene expression and cell structure.
However, while the expanded morphological and molecular features of astrocytes are compelling, we need to consider that direct experimental evidence linking these characteristics to cortical expansion remains limited. It is also plausible that the increased astrocytic complexity observed in primates arises as a response to the heightened structural and functional demands of more complex neuronal networks. Given that astrocyte morphology and identity are highly responsive to local cues, their diversity may reflect an adaptive capacity to support region-specific neuronal activity. Indeed, astrocytes have been suggested to retain a degree of plasticity, allowing them to dynamically adjust their molecular, structural, and functional properties in accordance with their microenvironment (Zheng et al., 2022).
Conclusion
The evolutionary expansion and increased complexity of the neocortex are deeply rooted in the specialization of its progenitor and glial cell populations. RG, particularly bRG, have evolved distinct morphological features and regulatory mechanisms—shaped by both genetic innovations and epigenetic modulation—that support their enhanced proliferative capacity. These changes not only increased neuronal output but also contributed to structural adaptations such as cortical folding. Astrocytes, likewise, have undergone remarkable evolutionary refinements, acquiring greater morphological complexity, transcriptomic heterogeneity, and functional versatility in primates and especially humans. From supporting glymphatic flow in gyrencephalic brains to sex-specific roles in circuit maturation and the emergence of specific astrocytic morphological traits, they feature as central players in both structural and functional brain evolution.
Integrating findings from both RG and astrocyte biology reveals an evolutionarily coordinated system that scaffolds the construction and maintenance of the expanded neocortex, beyond neurons. This cellular synergy potentially played a foundational role in the development of the large, complex, and metabolically demanding primate brain. As such, the evolutionary innovations in RG and astrocytes offer critical insight into the cellular architecture that underpins human cognition and highlight potential origins of our species' vulnerability to neurodevelopmental and psychiatric disorders.
Footnotes
- Received May 28, 2025.
- Revision received August 2, 2025.
- Accepted August 8, 2025.
-
The authors declare no competing financial interests.
- Correspondence should be addressed to Nereo Kalebic at nereo.kalebic{at}fht.org or Carmen Falcone at cfalcone{at}towson.edu.
- Copyright © 2025 the authors
References
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
